Article
Chemistry, Multidisciplinary
Elena D'Amato, Constantino Carlos Reyes-Aldasoro, Arianna Consiglio, Gabriele D'Amato, Maria Felicia Faienza, Marcella Zollino
Summary: This work presents a non-invasive, automated software framework that discriminates individuals with Pitt-Hopkins syndrome (PTHS) from healthy ones through analyzing morphological facial features. The software achieved an accuracy rate of 91% in classifying individuals with PTHS, compared to 74% recognition rate by pediatricians. Two geometric features related to the nose and mouth showed significant statistical differences between the two populations.
APPLIED SCIENCES-BASEL
(2021)
Article
Engineering, Electrical & Electronic
Ying Yi, Hu Zhang, Wei Zhang, Yahua Yuan, Changping Li
Summary: Working under fatigue states is inefficient and poses safety and health risks. This article introduces a novel fatigue detection algorithm based on facial multifeature fusion, showcasing its immediacy and accuracy. By processing video frames and extracting facial features in real-time, the algorithm can identify fatigue grades with high accuracy and quick response, providing reliable results in detecting fatigue behaviors.
IEEE SENSORS JOURNAL
(2023)
Article
Computer Science, Artificial Intelligence
Dan Li, Xin Zhang, Xiaofan Liu, Zhicheng Ma, Baolong Zhang
Summary: This paper proposes a fatigue detection method based on integrated facial features and Gate Recurrent Unit (GRU) judgment neural network, which uses neural networks and multi-task convolutional neural network to extract facial features and judge fatigue status. The real-time detection accuracy of this method can reach 97.47%.
JOURNAL OF REAL-TIME IMAGE PROCESSING
(2023)
Article
Engineering, Electrical & Electronic
Ying Yi, Hu Zhang, Wei Zhang, Yahua Yuan, Changping Li
Summary: This article presents a novel fatigue detection algorithm based on facial multifeature fusion, which exhibits promising properties of immediacy and accuracy, avoiding the inefficiency, safety concerns, and health problems caused by working under fatigue states.
IEEE SENSORS JOURNAL
(2023)
Article
Geochemistry & Geophysics
Zengfu Hou, Wei Li, Lu Li, Ran Tao, Qian Du
Summary: A novel method using multiple morphological profiles (MMPs) is proposed for hyperspectral change detection to make full use of spatial information. Experimental results demonstrate that the proposed detector achieves better detection performance on four real hyperspectral datasets.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Geochemistry & Geophysics
Mingjing Zhao, Lu Li, Wei Li, Ran Tao, Liwei Li, Wenjuan Zhang
Summary: The proposed method using multiple morphological profiles (MMP) can effectively detect various types of targets with different brightness in infrared images. By applying different attributes for feature extraction and a fusion strategy of different pruning values, the detection performance and robustness are improved.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
Article
Computer Science, Information Systems
Hao Yang, Li Liu, Weidong Min, Xiaosong Yang, Xin Xiong
Summary: Driver fatigue is identified as the main cause of traffic accidents. This study proposes a novel approach for detecting yawning based on subtle facial action recognition, using a 3D deep learning network and a keyframe selection algorithm to improve detection accuracy.
IEEE TRANSACTIONS ON MULTIMEDIA
(2021)
Article
Engineering, Biomedical
Neha, H. K. Sardana, R. Kanawade, N. Dogra
Summary: Photoplethysmography (PPG) is a non-invasive optical technique used for detecting cardiovascular diseases. Researchers propose a new set of morphological features for automated detection of multiple arrhythmias using rule-based and statistical learning-based approaches. The proposed methods are implemented and validated on retrospective and prospective datasets, and show comparable accuracy rates of 98.43%/94.16% (retrospective) and 94.16%/93% (prospective) for rule-based and statistical learning approaches, respectively.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Engineering, Civil
Guanglong Du, Tao Li, Chunquan Li, Peter X. Liu, Di Li
Summary: This paper proposes a method that uses a single RGB-D camera to extract three fatigue features and improves the accuracy of driver fatigue detection through a novel multimodal fusion recurrent neural network. By addressing the fuzziness and noise of the heart rate feature and identifying the relationship between features, the proposed method outperforms similar methods in both simulation and field experiments.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2021)
Article
Biology
Xiaolong Wu, Jianhong Yang, Yongcong Shao, Xuewei Chen
Summary: This study proposes a more sensitive mental fatigue assessment method based on an arbitrary channel EEG, using a combination of mathematical morphology and LSTM-CNN architecture. The results show that mathematical morphology methods can better reflect mental fatigue after sleep deprivation, and the LSTM-CNN architecture has strong generalization ability in dealing with different EEG channels.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Chemistry, Multidisciplinary
Tianjun Zhu, Chuang Zhang, Tunglung Wu, Zhuang Ouyang, Houzhi Li, Xiaoxiang Na, Jianguo Liang, Weihao Li
Summary: This research proposes a real-time comprehensive driver fatigue detection algorithm based on facial landmarks. By calculating eye and mouth features and establishing an assessment model, the fatigue status of drivers can be accurately and quickly evaluated.
APPLIED SCIENCES-BASEL
(2022)
Article
Mathematical & Computational Biology
Deguang Li, Zhanyou Cui, Fukang Cao, Gaoxiang Cui, Jiaquan Shen, Yongxin Zhang
Summary: This article presents an approach for assessing the state of online learning by combining blink detection, yawn detection, and head pose estimation. Experimental results show that this approach effectively evaluates the state of online learning and provides support for online education.
COMPUTATIONAL INTELLIGENCE AND NEUROSCIENCE
(2022)
Article
Engineering, Electrical & Electronic
Ying Yi, Zhushijie Zhou, Wei Zhang, Mingyue Zhou, Yahua Yuan, Changping Li
Summary: This article proposes a novel fatigue detection algorithm using the Dlib toolkit to mark facial feature points and considers the eye's multifeature. The algorithm achieves an optimal weight distribution for fusion and improves the reliability and error tolerance rate of detection. The testing results demonstrate a high accuracy rate of 95%, indicating strong potential in fatigue detection applications.
IEEE SENSORS JOURNAL
(2023)
Article
Computer Science, Artificial Intelligence
Muhammad Irfan Yousuf, Izza Anwer, Ayesha Riasat, Khawaja Tahir Zia, Suhyun Kim
Summary: The researchers propose a static malware detection system that can detect Portable Executable (PE) malware in Windows environment with high accuracy. By collecting malware samples and extracting relevant information, they combine machine learning, ensemble learning, and dimensionality reduction techniques to construct a system with a detection rate of 99.5% and an error rate of only 0.47%.
PEERJ COMPUTER SCIENCE
(2023)
Article
Multidisciplinary Sciences
Cangyan Xiao, Liu Han, Shuzhao Chen
Summary: A fatigue detection method for drivers based on face recognition under a single sample condition is proposed, with improvements in camera calibration, image enhancement, and detection efficiency through a symmetric algorithm. Experimental results show good overall detection effect, with higher accuracy and efficiency compared to traditional methods.
Article
Biochemistry & Molecular Biology
Zhiqiang Sha, Daan van Rooij, Evdokia Anagnostou, Celso Arango, Guillaume Auzias, Marlene Behrmann, Boris Bernhardt, Sven Bolte, Geraldo F. Busatto, Sara Calderoni, Rosa Calvo, Eileen Daly, Christine Deruelle, Meiyu Duan, Fabio Luis Souza Duran, Sarah Durston, Christine Ecker, Stefan Ehrlich, Damien Fair, Jennifer Fedor, Jacqueline Fitzgerald, Dorothea L. Floris, Barbara Franke, Christine M. Freitag, Louise Gallagher, David C. Glahn, Shlomi Haar, Liesbeth Hoekstra, Neda Jahanshad, Maria Jalbrzikowski, Joost Janssen, Joseph A. King, Luisa Lazaro, Beatriz Luna, Jane McGrath, Sarah E. Medland, Filippo Muratori, Declan G. M. Murphy, Janina Neufeld, Kirsten O'Hearn, Bob Oranje, Mara Parellada, Jose C. Pariente, Merel C. Postema, Karl Lundin Remnelius, Alessandra Retico, Pedro Gomes Penteado Rosa, Katya Rubia, Devon Shook, Kristiina Tammimies, Margot J. Taylor, Michela Tosetti, Gregory L. Wallace, Fengfeng Zhou, Paul M. Thompson, Simon E. Fisher, Jan K. Buitelaar, Clyde Francks
Summary: Small average differences in the left-right asymmetry of cerebral cortical thickness have been observed in individuals with autism spectrum disorder (ASD). This study used a large sample of data to examine the structural network effects of these regional alterations. The findings suggest that altered asymmetrical brain development in ASD may affect networks involved in executive functions, language-related processes, and sensorimotor processes.
MOLECULAR PSYCHIATRY
(2022)
Article
Nutrition & Dietetics
Heyu Meng, Yueying Wang, Jianjun Ruan, Yanqiu Chen, Xue Wang, Fengfeng Zhou, Fanbo Meng
Summary: The study indicates a relationship between the concentration of iron ions in peripheral blood and coronary atherosclerosis. Lower levels of iron ions in peripheral blood can serve as a predictive biomarker for coronary atherosclerosis.
Article
Chemistry, Multidisciplinary
Daiguo Deng, Zengrong Lei, Xiaobin Hong, Ruochi Zhang, Fengfeng Zhou
Summary: This study used a heterogeneous graph neural network (MolHGT) to represent molecular structures, which showed improved performance in molecular property predictions compared to existing studies.
Article
Biochemical Research Methods
Yaqi Zhang, Gancheng Zhu, Kewei Li, Fei Li, Lan Huang, Meiyu Duan, Fengfeng Zhou
Summary: This study presents a novel HLAB feature engineering algorithm for detecting HLA-I binding peptides using natural language processing and deep neural networks. The experimental results show that the proposed algorithm outperforms existing methods in predicting peptides binding to specific HLA alleles, achieving the best performance in most prediction tasks.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Biochemical Research Methods
Qian Wang, Meiyu Duan, Yusi Fan, Shuai Liu, Yanjiao Ren, Lan Huang, Fengfeng Zhou
Summary: This study focused on the classification problem in the context of large feature space and small sample space in OMIC data. The researcher utilized a Siamese convolutional network to transform the OMIC features and achieved improved classification accuracies for binary classification problems.
JOURNAL OF BIOINFORMATICS AND COMPUTATIONAL BIOLOGY
(2022)
Article
Medicine, Research & Experimental
Chengfeng Xu, Ruochi Zhang, Meiyu Duan, Yongming Zhou, Jizhang Bao, Hao Lu, Jie Wang, Minghui Hu, Zhaoyang Hu, Fengfeng Zhou, Wenwei Zhu
Summary: This study used RNA sequencing to detect biomarkers for immune thrombocytopenia (ITP) and provided a highly accurate ITP detection model. The biomarkers suggested that ITP onset may be associated with various transcribed components. This research provides important insights for the diagnosis and investigation of ITP.
MOLECULAR THERAPY-NUCLEIC ACIDS
(2022)
Editorial Material
Genetics & Heredity
William C. Cho, Fengfeng Zhou, Jie Li, Lin Hua, Feng Liu
FRONTIERS IN GENETICS
(2022)
Article
Dermatology
Shuai Liu, Yusi Fan, Meiyu Duan, Yueying Wang, Guoxiong Su, Yanjiao Ren, Lan Huang, Fengfeng Zhou
Summary: Acne is a common skin lesion in adolescents, and severe or inflammatory acne can lead to scarring, affecting patients' quality of life and job prospects. Diagnosis of acne is done by counting the number of lesions, but this can be a labor-intensive task prone to errors, making the development of automatic diagnosis methods important.
SKIN RESEARCH AND TECHNOLOGY
(2022)
Article
Engineering, Electrical & Electronic
Xin Feng, Yingrui Li, Zhang Hang, Zhang Fan, Qiong Yu, Ruihao Xin
Summary: This paper proposes a joint learning text information extraction method TBR-NER based on topic recognition and named entity recognition, which is verified to be reliable and effective in experiments and helpful for epidemic prediction, prevention, and control.
JOURNAL OF SENSORS
(2022)
Article
Biology
Meiyu Duan, Yueying Wang, Ya Qiao, Yangyang Wang, Xingyuan Pan, Zhuyu Hu, Yanyue Ran, Xian Fu, Yusi Fan, Lan Huang, Fengfeng Zhou
Summary: The transcriptome is a comprehensive reflection of gene expression in a sample. This study proposes a novel approach to understanding the regulatory relationships between transcription factors (TFs) and their target genes (mRNAs) by quantitatively analyzing the differences in their expression levels. By using a multi-input multi-output regression model, this study explores the quantitative transcription regulation relationships of metabolism-related genes. The findings suggest that certain genes can exhibit differential regulation patterns even when their expression levels do not differ significantly, and these "dark biomarkers" may have important implications for disease detection and classification.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Computer Science, Information Systems
Changfan Luo, Yiping Xu, Yongkang Shao, Zihan Wang, Jianzheng Hu, Jiawei Yuan, Yuchen Liu, Meiyu Duan, Lan Huang, Fengfeng Zhou
Summary: Feature engineering method EvaGoNet is proposed in this study, which improves the decoder module of the Gaussian mixture variational autoencoder using the Wasserstein generative adversarial network with gradient penalty (WGANgp) and updates the latent features by embedding the top-ranked original features based on their discriminative powers. Comprehensive experiments demonstrate that EvaGoNet-encoded features outperform existing classifiers on 12 benchmark datasets, especially on small, imbalanced datasets col (accuracy = 0.8581), spe (accuracy = 1.0000), and leu (accuracy = 0.8021). EvaGoNet-engineered features enhance binary classification outcomes on six high-dimensional, imbalanced bioOMIC datasets. EvaGoNet achieves a medium-ranked training speed among compared algorithms and exhibits considerably fast prediction speeds in the testing sample predictions. Therefore, EvaGoNet can serve as a candidate feature engineering framework for practical applications requiring one training procedure and multiple prediction tasks of testing samples.
INFORMATION SCIENCES
(2023)
Article
Pharmacology & Pharmacy
Ruihao Xin, Xin Feng, Hang Zhang, Yueying Wang, Meiyu Duan, Tunyang Xie, Lin Dong, Qiong Yu, Lan Huang, Fengfeng Zhou
Summary: This study used a feature-engineering approach model to quantitatively measure the correlation between mRNA and transcription factors in CLL. The analysis found seven mRNAs with significantly differential model-based transcription regulation values, which were called "dark biomarkers" as their expression levels did not show differential changes in CLL patients. The overlapping lncRNAs may contribute to the expression miscalculations of these dark biomarkers.
PERSONALIZED MEDICINE
(2023)
Article
Engineering, Electrical & Electronic
Ruihao Xin, Fengbo Miao, Ping Cong, Fan Zhang, Yongxian Xin, Xin Feng
Summary: Emotion recognition is crucial for computers to understand human emotions. Traditional EEG emotion recognition methods have limitations, and to improve accuracy, a multi-view feature fusion attention convolutional recurrent neural network (multi-aCRNN) model is proposed. Multi-aCRNN combines CNN, GRU, and attention mechanisms to deeply fuse features from multiple perspectives. It achieves high classification accuracy in arousal and valence tasks. Overall, multi-aCRNN can effectively integrate EEG features and improve emotion recognition classification.
JOURNAL OF SENSORS
(2023)
Article
Engineering, Electrical & Electronic
Ruihao Xin, Xin Feng, Tiantian Wang, Fengbo Miao, Cuinan Yu
Summary: In this study, a deep multitask-based multiscale feature fusion network model (MEAT) is proposed to address the limitations and poor adaptability of traditional convolutional neural network models for complex jobs. The model achieved an accuracy of 99.95% for the total task of fault six classification and accurately diagnosed bearing faults by considering fault size and fault type through multi-task mapping decomposition.
Article
Chemistry, Medicinal
Ruihao Xin, Fan Zhang, Jiaxin Zheng, Yangyi Zhang, Cuinan Yu, Xin Feng
Summary: Choosing appropriate encoding methods is challenging in tasks related to DNA sequence classification. We introduced a new trainable ensemble method based on the attention mechanism SDBA, which allows the model to dynamically ensemble features from different perspectives and acquire and weight sequence features voluntarily.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)